I use the svm function (for regression) to make forecast like I would with for exemple the arima function:
which returns different attributes :

  1. prediction$mean which is the actual prediction
  2. prediction$lower and prediction$upper which are the boundaries of the confidence intervals on each points of the prediction$mean.

I would like the svm function (from e1071 package) to return a more detailed answer than just the value (like the forecast() would).
But I guess it is not implemented in the function yet. Is there another function to do it ? Or should I use bootstrap methods to try to estimate those boundaries? And if I should use this are they pre-implemented version of them instead of using sample over a for loop which is very time-consuming ?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.